In spite of dramatic improvements in cure rate for the most common childhood malignancy, acute lymphoblastic leukemia (ALL), one in four to five children will suffer disease reoccurrence and their prognosis is dismal. In fact relapsed ALL is the most common cause of non-accidental death in childhood. Attempts to improve relapse outcome by intensifying chemotherapy including the use of stem cell transplantation have failed to cure the majority of children. While many prognostic factors can be used to stratify therapy, only immunophenotype (T-ALL inferior prognosis), site of relapse (bone marrow inferior to isolated extramedullary) and timing of relapse (early relapse defined as <36 months from initial diagnosis inferior compared to late relapse) are predictive of salvage. To discover the underlying pathways that mediate drug resistance in childhood ALL we have used high throughput genomic techniques to analyze differences in global gene expression and copy number abnormalities (CNAs) in matched diagnosis-relapse paired samples from children enrolled on Children's Oncology Group (COG) protocols. Our early results indicate that most cases of early relapse are characterized by a proliferative gene signature and we suggest that these clones exist at diagnosis. In contrast, we hypothesize that many cases of late relapse occur through acquisition of additional genetic and/or epigenetic changes. Importantly, we have identified attractive targets for novel therapeutic approaches including a subset that have been validated in preclinical assays. The goals of this application are to extend these observations and confirm our hypothesis through the following specific aims: 1) discover biological pathways responsible for the emergence of resistant disease by identifying gene signatures associated with early and late relapse using oligonucleotide arrays, 2) to determine the critical lesions that drive resistance by identifying CNAs unique to the relapsed clone using Affymetrix 6.0 SNP arrays, 3) to validate the critical roles of the pathways identified in aims 1 and 2 by modulating expression in a panel of ALL cell lines exposed to chemotherapy and to establish whether transcript levels correlate with the drug sensitivity of blasts exposed ex vivo to agents used in ALL treatment, and 4) to formally test our hypothesis by documenting the evolution of gene expression signatures in residual blasts at end induction, and by "backtracking" relapse specific deletions to determine if relapsed clones were present at diagnosis before the application of therapy. The results of these studies will lead to an understanding of the cellular origin of relapsed disease and the pathways that are responsible for treatment failure. Such information will lead to the discovery of pathways that can be targeted in future trials. PUBLIC HEALTH RELEVANCE: Although the majority of children who are newly diagnosed with acute lymphoblastic leukemia (ALL) are now cured, recurrences develop, often unpredictably, in 25% of cases. Success in treating relapse has significantly lagged behind that of newly diagnosed disease as evidenced by the fact that only approximately one third of children with recurrent ALL survive long-term. The goal of this project is to develop a better understanding of the molecular genetic events that lead to relapse so that this knowledge can be used to identify more effective preventative and therapeutic strategies.